beamforming with acoustic phased array

[dgsharp]

I'm working on a very crude acoustic phased array for fun, to help me learn about beamforming. I'm shooting for very low cost and as simple as possible, and I was hoping I could get some advice. (Sorry if this is the wrong forum, I couldn't decide which was most appropriate! I'm using a microcontroller but the phenomena are analog.) Here's what I have right now:

- 10 piezo transducers of the type with a disk in the back of a plastic housing, and the sound emanates from a hole in the center. Resonant frequency seems to be around 3.8kHz
- low-power microcontroller (Arduino in this case)
- element spacing is 1.3", I can change this without too much trouble

The way I'm driving the transducers is super cheesy, directly from the microcontroller GPIOs. I actually just toggle the output state once (high/low -- 0/5V) and the transducer rings at its resonant frequency for around 5ms or so. I can successfully steer a beam in any azimuth angle from about -60 to 60 degrees, but I'm only getting maybe 4dB increase in SPL along the axis of the beam. So I would say it is without a doubt working, but I was really hoping that it would be much less audible for off-axis directions. Does anyone have any ideas? Running it continuously rather than single "chirps" doesn't seem to change this.

Here were some of my thoughts:

- Maybe the element spacing is more important than I thought? I somewhat arbitrarily chose 1.3" spacing just to make it a reasonable size. Do you think going to a lambda/2 spacing (about 1.7") would make it any quieter off the beam axis, or maybe just louder on-axis?
- Maybe I need to filter my square wave signal that is driving the transducers? I could try an RC filter. I have some doubts though since the transducer does most of the ringing on its own and the waveform looks quite sinusoidal to my eye, I just give it one little "kick" usually.

Final question: if I had a basic simulator it might help me play with factors like spacing and see how to optimize my system. Any recommendations?

Thanks so much for any help you can give! I am hoping to make a simple scanning rangefinder demo for fun.

 

[godfreyl]

The transducers won't all ring at exactly the same frequency, so it would be better to drive them continuously. Then they will at least all be emitting the same frequency.

The next problem is that their frequency responses may not all be the same, in which case the amplitude and phase of the outputs may be different. That would be more difficult to sort out.

 

[dgsharp]

Thanks for the ideas, I'll think about that. I have driven them continuously and didn't find an audible difference to my ear. Maybe that in combination with a filter to attenuate the odd harmonics would help some...